Plasma phosphorus-doped oxides

J. Houskova, K.N. Ho, M.K. Balazs, "Characterization of Components in Plasma Phosphorus Doped Oxides", ACS Symposium on Materials Characterization in Microelectronics Processing (ACS National Meeting), St. Louis, 1984. 

To Identify the foul smelling gas, a closed system was built in which the film was dissolved. The liberated gas was driven to a reaction chamber and oxidized. If the gas were phosphine, it would be converted to phosphate ion and as such could be analyzed. This test was positive for every plasma and plasma enhanced phosphorus doped oxide that was analyzed. Deductively it was concluded that the gas... 

Since Si02 substrates appear frequently during IC fabrication, the adhesion test results for this substrate are important. Four types of oxides have been extensively tested. They are (1) thermal oxide grown at 7>1000°C, (2) softer oxide processed by conventional spin-on-glass technology, (3) phosphorus-doped LPCVD oxide, and (4) low-temperature (200°C) plasma deposited oxide. 

The immediate concern about the presence of phosphine in doped oxide films was safety in handling these materials. The permissible exposure limit (PEL) set by OSHA is 0.3 ppm, and the concentration immediately dangerous to life and health (IDLH) is 200 ppm. All precautions must be taken when etching plasma PSG of BPSG wafers. A basket of twenty 3 inch wafers, 5000 angstroms thick, containing 5 wt. % phosphorus could release as much as 2500 ppm of phosphine or 12.5 times the IDLH. 

It is also well known (2) that not all oxoacids of phosphorus react with molybdenum ion to form a positive blue complex at the same rate. Time studies were done on plasma PSG wafers. Figure 1 shows distinct differences between the phosphate ion and doped oxide film from a wafer when analyzed with and without oxidation. 

Table III. Weight Percent Phosphorus in Plasma Doped Oxides...

The determination of specific phosphorus compounds in thin films is important. Only through wet chemical analysis was it possible to first discover the presence and then to accurately measure the quantities of P2Os, P203, and phosphine found in plasma, plasma-enhanced, LPO-LTO (low-pressure oxide-low-temperature oxide), and CVD (chemical vapor deposition) processes (3). Methods such as X-ray or FTIR spectroscopy would have seen all phosphorus atoms and would have characterized them as totally useful phosphorus. In plasma and plasma-enhanced CVD films, phosphine is totally useless in doping processes. 

If desired, plasma oxide films can be doped much as the plasma nitride film we discussed earlier. In fact, doping with boron and phosphorus has been carried out as an alternative to the standard atmospheric-pressure thermal CVD process for BPSG.11 12 The latter process has the drawbacks of high defect density and poor thickness uniformity, so it was hoped that plasma BPSG would be an improvement. However, there are differences in the films in terms of H2 and N2 content, and their effect on reflow temperature, intrinsic stress and passivation effectiveness had to be examined. 

---